Skip to main content

Using the Assistant

Using the Default Copilot Assistant​

  1. Access the Assistant Page

    • Upon entering the assistant page, you will see "Copilot" as the default assistant option.
    • This default assistant comes with deep thinking capabilities and is configured with a default model.

  2. Select Model and Auto-Matching

    • During regular use, you can select a suitable model based on your needs or directly choose "Auto" to let the system automatically match the most appropriate model.

  3. Online Search and Other Interactive Features

    • The page supports enabling the online search feature, helping you find answers from the web.
    • If you need to upload images or use voice input interaction, these features are also directly available.

💡 Tip: Image uploads only support png, jpg, and jpeg formats.

  1. Assistant Conversation Interface
    • At the top of the assistant conversation interface, there are several practical features:
      • Click the corresponding entry to start a new conversation.
      • If you need to review the chat history with the assistant, a dedicated entry allows you to view it.
      • You can select interaction language and switch languages as needed.
      • To clear the current conversation content, use the clear context feature.
      • You can also maximize the conversation window to enhance the user experience.

  1. History and Frequently Used Assistants
    • The left side of the interface features a history view, allowing you to review all assistant conversation histories.
    • You can rename or delete conversation histories.
    • Frequently used assistants are displayed at the top of the interface for quick access.
    • Click "More Assistants" to jump to the Assistant Plaza and select other assistants.

  1. Q&A Interaction
    • Copy response content: You can copy all content provided by the assistant.
    • Like responses: If you are satisfied with the assistant's response, you can like it.
    • Dislike responses: If you are dissatisfied with the assistant's response, you can dislike it.
    • QnA correction: Select the QnA in the enterprise space, input the question and answer, and click confirm to make corrections.

Question Guidance and Follow-Up​

  1. Intelligent Question Recommendations and Auto-Completion

    • When users input questions in the conversation box, the system uses intelligent algorithms to automatically generate related questions that may interest the user for reference.
    • If the user's question is incomplete, the system will automatically complete it, ensuring a smoother and more efficient interaction process.

  2. Follow-Up Functionality

    • After Q&A, users can ask follow-up questions based on the assistant's response, files, or tags.
    • The specific method for follow-up is:
      • Click on the file or tag provided in the assistant's response.
      • Click "Re-ask" below the response, and the assistant will narrow the scope based on the file and tag for further answers.

Private Knowledge Q&A​

When using the assistant, if you want to conduct Q&A based on a private knowledge base, you can configure it as follows:

  • Go to the assistant configuration page and disable the "Allow File Upload" button under Knowledge Source Configuration.
  • Click the "+" button on the right side of the knowledge source to add private knowledge base content to the assistant.
  • Click the settings button on the right side of the knowledge source and enable the "Enforce Private File Q&A" option.
  • After configuration, click the "Save Configuration" button in the upper right corner to successfully create a private Q&A assistant.
  • After successful configuration, you can conduct tests to ensure the private Q&A functionality operates normally.
    💡 Tip: When the assistant enforces private Q&A, the online search skill will be disabled by default, and online searches will no longer be possible.

Single/Multiple Document Q&A​

Single Document Q&A​

If you want to conduct Q&A based on a single document while using the assistant, you can select a single document in the knowledge scope after configuring the knowledge source.

  • Click "Knowledge Scope", select a single document, and click "Confirm".
  • Once a single document is selected, you can conduct Q&A specific to that document.

In addition to selecting a single document in the knowledge scope for Q&A, you can also use AI summarization for document Q&A.

  • Find and preview the document in the knowledge base.
  • On the preview page, click "Conversation" to open the assistant named "File Copilot" for document Q&A.

Multiple Document Q&A​

Multiple document Q&A is similar to single document Q&A. After configuring the knowledge source, select multiple documents in the knowledge scope.

  • Click "Knowledge Scope", select multiple documents, and click "Confirm".
  • Once multiple documents are selected, you can conduct Q&A specific to these documents.

tip
  • When selecting files in the knowledge scope, you can also directly select the entire knowledge base for Q&A, provided the knowledge base is selected in the assistant configuration.
  • Additionally, the knowledge scope does not support selecting folders; only files or knowledge bases can be selected.

Knowledge Base Q&A​

There are two ways to conduct knowledge base Q&A:

  • One is to configure the knowledge base in the assistant, and select the knowledge base in the knowledge scope during Q&A.

  • The other is to enter the knowledge base page, where a "Expand Q&A Chat" button is located in the upper right corner. Click it to use the knowledge base Q&A assistant.

Assistant Plaza​

All assistants are located in the Assistant Plaza, where you can search for any assistant. If you have permissions, you can also create assistants in the plaza.

The entry to the Assistant Plaza is located under the Frequently Used Assistants List. Click "More Assistants" to access the plaza.

How to Create an Assistant?​

There are three ways to create an assistant: Basic Workflow, Advanced Workflow, and Share Code Creation.

Among them, the basic workflow is more suitable for novice users, the advanced workflow allows assistants to achieve more complex functions, and share code creation enables direct migration of existing assistants.

💡 Tip: Creating an assistant requires assistant creation permissions. Please ensure you have this permission before proceeding.

Assistant Creation Entry:

  • Assistant Plaza -> Create Assistant

  • Settings -> Agent & Copilot -> Assistant Management -> Create Assistant

Basic Workflow​

Steps for creating an assistant using the basic workflow:

  • Click "Create Assistant" from the assistant creation entry, then select "Basic Workflow Creation".

  • Enter the assistant name, select an assistant avatar, choose a model group, select an assistant category, and add a description.

    • Assistant Name: Enter the name of the assistant as its identifier.
    • Assistant Avatar: Select the default avatar for the assistant (currently, uploading avatars is not supported).
    • Model Group: Configure a suitable model group for the assistant.
    • Assistant Description: Provide a brief description explaining the assistant's functionality and purpose.
    • Assistant Category: Choose the group where the new assistant will be located (multiple selections are allowed).

  • Click "Create". After creation, you will enter the basic workflow assistant configuration page. Configure and publish the assistant for use.

How to Configure a Basic Workflow Assistant?​

  • There are two ways to access assistant configuration:

    • Frequently Used Assistants List: Click the "..." next to the assistant name and select "Configure".
    • Path: Management -> Assistant Management -> Assistant Management, then select the assistant to configure.

  • Prompt: Enter the assistant's prompt, or use intelligent generation for existing prompts. Prompt length is limited to 2000 characters.

  • Opening Statement: Enter the assistant's opening statement, or use intelligent generation based on prompts or existing opening statements. Opening statement length is limited to 2000 characters.

  • Model Group: Click "+" to add a model group. Multiple models are supported.

  • Skills: Click "+" to add one or more skills, or add recommended skills.

  • Conversation Experience:

    • Mask Collection: Click "+" to add a mask collection.
    • Conversation Settings: Enable settings such as "User Question Suggestions, Question Guidance, Chat History, Conversation Feedback, Keyword Review," etc.

  • Knowledge Base:

    • Knowledge Base: Click "+" to add a knowledge base.
    • Knowledge Base Configuration: Modify detailed settings such as "Retrieval Strategy, Private Q&A, Retrieval Method," etc.

  • Data Source: Click "+" to add a data source as the assistant's Q&A data source.

  • Assistant Preview and Debugging: The right side of the interface allows you to preview and test the assistant. Save the configuration after testing.

Model Group​

Model groups must first be added by administrators in system management. Multiple models can be added to the same model group, which can then be configured for the assistant.

Model group addition path: Settings -> System Management -> Model Management -> Model Group -> Create Model Group

  • Click "Create Model Group".
  • Enter the model group name.
  • Select the models to include in the group (multiple selections allowed).
  • Choose whether to enable adaptive model deployment.
  • Choose whether to enable deep thinking models.
  • Click "Save".

Skills​

Currently available skills include Text-to-Image, Online Search, API Function Call V2, News Query Tool, Python Function, and Webpage Reading.

  • Text-to-Image: Generates images based on input text content.
  • Online Search: Connects to the internet to retrieve publicly available content (e.g., news, information).
  • API Function Call V2: Calls the second version of application program interface functions for specific data interactions or operations.
  • News Query Tool: A specialized tool for searching and retrieving various news information.
  • Python Function: Functions written in Python for encapsulating code logic, data processing, algorithm computation, and specific program tasks.
  • Webpage Reading: Extracts text, data, and other content from webpages to parse webpage information.

Knowledge Source​

Allow File Upload:

  • Enabling file upload prevents adding knowledge base content as a knowledge source.
  • Disabling file upload allows selective addition of personal or enterprise space knowledge bases as knowledge sources.

Knowledge Base Configuration:

  • Retrieval Strategy: Hybrid Search, Embedding Search, Text Search

    • Hybrid Search: Combines vector and full-text search results, returning reordered results.
    • Embedding Search: Finds fragments based on similarity, with some cross-language generalization capabilities.
    • Text Search: Finds fragments based on keywords, suitable for retrieving specific keyword or noun fragments.

  • Maximum Recall Quantity: Range 1-8. Setting this too high or too low may affect file content retrieval. Recommended value: 3-5.

  • Metadata Filtering: None, Filter, Weight

  • Enforce Private File Q&A: When enabled, online search and other skills are disabled, and the assistant's responses are limited to knowledge base content.

  • Document Matching Similarity: Range 0-1. Higher similarity indicates greater thematic and textual similarity of recalled documents but may make it harder to find highly relevant documents. Recommended value: ~80%.

  • QnA Matching Similarity: Range 0-1. Similar to document matching similarity, higher values indicate greater matching accuracy but may make it harder to find QnAs. Recommended value: ~90%.

  • Display References: When enabled, the assistant lists referenced documents in its responses, enhancing credibility.

💡 Tip: Maximum recall quantity, document matching similarity, and QnA matching similarity should align with actual needs. If no special requirements exist, default settings are recommended.

Data Source​

Assistant data sources can match connected data sources, serving as reference material for the assistant.

Steps to add assistant data sources are simple: Click the "+" next to the data source, select the source, and click "Confirm".

Conversation Experience​

User Question Suggestions: After the assistant responds, it provides question suggestions based on the context.

Question Guidance: During user-assistant interaction, related question guidance is provided, leveraging model capabilities to predict and complete user questions.

Chat History: Whether to retain assistant chat history. If disabled, chat history will not be accessible.

Conversation Feedback: Users can interact with the assistant's responses by liking or disliking them, aiding response optimization.

Enable Keyword Review: At least one of input or output content review must be enabled.

Advanced Workflow​

Steps for creating an assistant using the advanced workflow:

  • Click "Create Assistant", then select "Advanced Workflow Creation".
  • Enter basic assistant information: Assistant Name, Avatar, Description, Category.
    • Assistant Name: Enter the name of the assistant as its identifier.
    • Assistant Avatar: Select the default avatar for the assistant (currently, uploading avatars is not supported).
    • Assistant Description: Provide a brief description explaining the assistant's functionality and purpose.
    • Assistant Category: Choose the group where the new assistant will be located.
  • After entering basic information, click "Create" to access the advanced workflow page.

How to Configure Advanced Workflow?​

Advanced Workflow Configuration:

  • Start, End: Built-in input/output modules. Input/output parameters and fields can be customized.
  • Model: Select the model to use in this module, input variables from other modules, and edit prompts and output messages, saving them as variables.
  • Skills: Select a skill to perform input/output actions through it.
  • Data Source: Choose a data source to add reference variables.
  • Code: Write and create custom code functions based on output variables from other modules.
  • Knowledge Base: Retrieve the most relevant information from the selected knowledge base based on input variables and return it.
  • Selector: Connect multiple downstream branches. If the set condition is met, only the corresponding branch runs; if none are met, the "Otherwise" branch runs.
  • Intent Recognition: Recognizes user input intent and matches it with preset intent options.
  • Text: Processes multiple character-type variables for formatting.

💡 Tip: For detailed creation, refer to "Advanced Workflow for Building Assistants and Applications."

Share Code Creation​

Share code creation can be understood as assistant migration. Its core principle is generating a unique share code for an existing assistant, allowing users to quickly create a new assistant by entering the code, achieving seamless functionality replication and dissemination.

Steps for creating an assistant using share code:

  • Enter the Assistant Plaza, click the three dots next to the assistant.
  • Select "Share" to generate a configuration code. Copy the code for later use.
  • Click "Create Assistant", then select "Share Code Creation".
  • Paste the copied configuration code into the code field and click "Create".
  • You can adjust the assistant created via share code in the configuration page to better suit the scenario.